Submitted to: Food Processing Automation Conference Proceedings
Publication Type: Proceedings
Publication Acceptance Date: June 16, 2008
Publication Date: June 28, 2008
Citation: Park, B., Kise, M., Lawrence, K.C., Windham, W.R., Yoon, S.C. 2008. Portable multispectral imageing instrument for food industry. Food Processing Automation Conference Proceedings. ASABE Publication Number 701P0508cd Interpretive Summary: This paper reports on a cost effective portable multispectral instrument that inspectors and plant personnel can use for quality control during their inspection process. Food safety in the poultry industry, especially in small plants, is an ongoing problem. Identification and separation of the birds contaminated by feces are very important to protect the consumer from a potential source of food poisoning. The current method of inspecting fecal contamination, however, is through human visual observation with the criteria of color, consistency, and composition used for identification. Therefore, there is a need to develop instrumental methods which can reduce inspector fatigue, variability, insure continuous inspection, and provide a safe poultry food supply for the consumer. The ARS poultry imaging lab in Athens, Georgia has patented a method for detecting contaminants on the surface of foods based on multi- and hyperspectral imaging techniques. Based on the patented method, a hand-held instrument for contamination detection was developed. The instrument uses only 2 or 3 wavelengths to detect fecal contamination. It has the potential to improve FSIS’s food safety inspection program in identifying, developing, and validating new technologies that are economically viable, especially for small poultry processing plants to help them meet food safety requirements.
Technical Abstract: The objective of this paper is to design and fabricate a hand-held multispectral instrument for real-time contaminant detection. Specifically, the protocol to develop a portable multispectral instrument including optical sensor design, fabrication, calibration, data collection, analysis and algorithm development for detecting poultry fecal contaminants in real-time was demonstrated. The imaging system was able to acquire multispectral images at up to three specified spectral bands simultaneously by utilizing identified interference filters, monochrome cameras, and optical components. Since the optical filters are interchangeable and adopted from off-the-shelf components, the multi-band spectral instruments can be easily assembled. Considering the field of view and lens distortion, the optimum optics were selected using two step image registration algorithms such as lens distortion correction and image projection. For two-band camera system, image calibration method provided accurate registered image of 3-D object with less than one pixel registration error for a 15 mm lens. However severe distortion occurred for a 10 mm lens system, which indicated that an accurate lens distortion correction, especially radial distortion correction is considerably needed to achieve precise image registration. The performance of prototype multispectral instruments can be improved with optimum design of optical components and advanced calibration protocol for the applications at various food processing industries.